The Study on Stock Selection Based on Deep Learning and the Billboard Data

The Billboard is a collection of active stocks in the Chinese stock market. It provides investors with a range of stocks to select from and shortens the time that the investors use to research the stocks. When stocks appear on the Billboard, they may generate positive or negative returns in the future. Therefore, it is an important topic to study how to select stocks from the Billboard which would generate positive returns in the future. In this paper, a hybrid model based on deep learning is proposed to select stocks from the Billboard. This model consists of three submodels for price prediction and movement prediction of stocks, and finally stock selection will be made based on the output of these three models. The empirical study shows that the model constructed in this paper can effectively make stock selection and achieve positive returns.